Support Vector Machines and Fuzzy Nonlinear Regression for Intelligent Identification of Urban VANET Constraints

Nowadays, video-on-demand (VoD) applications are becoming one of the tendencies driving vehicular network users. In this paper, considering the of ve-hicular network opens up to different types of communications in order to meet the needs of the wide variety of new applications envisaged within the framework of the Intelligent Transport System (ITS). In this work, we seek to establish a list of possibilistic concepts in order to efficiently identify the strict parameters of ur-ban VANET networks. To this end, we use linear optimization under constraints. We apply in parallel to this first proposition a minimization of a validated quadrat-ic criterion with the appearance of fuzzy least squares. To arrive at a quadratic resolution under constraints, different distances were managed and various con-straints were introduced in the optimization problem. We have shown that the da-ta independent criterion in urban VANETs can overcome the failure problem in terms of robustness. To assess the comparative effectiveness of our solutions, many experiments are carried out. The obtained results showed that the proposed identification scheme will allow an increase in the performance of Urban VANET networks with different load conditions.